The effects of information privacy and online shopping experience in e-commerce.

Bernard, Elena Kiryanova ; Makienko, Igor

INTRODUCTION

For a decade information privacy has been one of the central issues in e-commerce research across many disciplines. Extensive research has shown that due to a different nature of shopping environment, consumers perceive online transactions as risky, form heightened privacy concerns and such concerns become the main barrier for electronic commerce (Hoffma, Novak & Peralta, 1999). In marketing, information privacy has been linked to online trust (Bart et al., 2005; Eastlick et al., 2006; Hoffman et al., 1999; Pan & Zinkhan, 2006), e-service quality (Zeithaml et al., 2002), and online purchasing (Malhotra et al., 2004). Some researchers have examined the antecedents of e-shoppers' privacy perceptions, advocating various privacy management strategies such as opt-in/opt-out tactics, monetary compensation for customer information, and third-party privacy seals (Culnan, 1995; Goodwin, 1991; Rifon et al., 2005). In this study we will investigate yet another privacy management strategy that focuses on the transparency of the e-tailer's consumer information practices.

The supporters of the transparency strategy argue that customers would be more willing to trust the e-tailer with their personal information if the e-tailer explained the intended uses of customer information (Hoffman et al., 1999; Pan & Zinkhan, 2006). This view suggests that the mere transparency of the e-tailer's information practices can reduce customers' privacy concerns and enhance their perceptions of the e-tailer trustworthiness. However, research shows that many consumers either do not read or do not fully comprehend e-tailers' information privacy policies thus raising questions about their effectiveness in reducing customers' information privacy concerns (Cranor et al., 2006; Meinert et al., 2006; Milne & Culnan, 2004; Milne et al., 2006; Vail et al., 2008, Nehf, 2007; Proctor et al., 2008). Furthermore, little if anything is known about the effect of customers' previous online shopping experience on their reactions to the e-tailer's information privacy policy. Is it effective for all customers regardless of whether they are novice or experienced online shoppers? While Bart et al. (2005) show that both a consumer's Internet shopping experience and the website privacy policy have a positive influence on e-trust, we have not found any research that looked at the interaction between these two variables. This study will attempt to address these issues by specifying a structural equation model where customers' perceptions of e-tailer's information privacy policy, their online shopping experience, and the interaction of these two variables are explicitly linked to their privacy concerns and their perceptions of e-tailer trustworthiness.

THEORETICAL FRAMEWORK

This section is organized as follows. First, we will define online trust and explain conceptual underpinnings of e-tailer trustworthiness, which is the focal dependent variable in this study. Then we will discuss the hypothesized relationships among customers' perceptions of e-tailer information privacy policy, their online shopping experience, their privacy concerns, and their perceptions of e-tailer trustworthiness. Figure 1 depicts the research model.

[FIGURE 1 OMITTED]

Online Trust vs. Trustworthiness

Online trust has been defined in literature as a person's willingness to accept vulnerability based on positive expectations about the e-tailer's intentions and behaviors (Rousseau et al., 1998). These positive expectations encompass the customer's perceptions of the website's competence in performing required functions and his or her perceptions of the firm's good intention behind the "online storefront" (Bart et al., 2005). In other words, online trust is a behavioral outcome of a customer's belief in the e-tailer trustworthiness. It is important to distinguish between trust (behavior: i.e., willingness to depend) and perceived trustworthiness (cognition: i.e., beliefs about trustee's integrity, competence, and benevolence). Despite their conceptual differences, these two constructs have been often used interchangeably as the following definitions of trust show: the willingness of one party to be subject to risks brought by another party's actions (Gambetta, 1988); the belief that e-tailer will not "behave opportunistically by taking advantage of the situation" (Gefen et al., 2003, p. 54); a belief that the seller has integrity, competence, and benevolence (Bhattacherjee, 2002; Doney & Canon, 1997; McKnight et al., 2002) and one's willingness to accept vulnerability based on positive expectations about the other party's intentions or behaviours (Rousseau et al., 1998). Yet research shows that the three trustworthiness dimensions (integrity, competence, and benevolence) have different behavioral outcomes, making a plausible case for separating trust and trustworthiness (Gefen et al., 2003). In our study, we will examine e-tailer trustworthiness, defined in terms of the e-tailer's dependability, competence, integrity, and responsiveness. This choice of construct is justified by the purpose of the study, which is to examine the changes in customers' perceptions of an e-tailer resulting from the e-tailer's privacy policy strategy and customers' general experience with shopping on the Internet.

The Effects of Information Privacy Policy

Consistent with the principles of the social contract theory (Dunfee et al., 1999), consumers enter into a social contract with a company every time they provide their personal information (Culnan, 1995; Milne & Gordon, 1993). In exchange, they expect the firm to uphold their rights to limit the accessibility and to control the release of their personal information. The company's failure to fulfil its obligations in regard to customers' information privacy results in a breach of social contract and erosion of trust. Consequently, social contract theory suggests that consumers' decision to purchase from an online firm depends on their perceptions of the firm's privacy practices, which can be gleaned from its privacy policy statement.

Privacy disclosures posted on the e-tailer's website may have both direct and indirect effects on consumers' perceptions of e-tailer trustworthiness. On the one hand, the information provided in these disclosures may address specific consumer concerns in regard to the firm's handling of their personal information thus resulting in lower privacy concerns (Campbell, 1997; Gengler & Leszczyc, 1997; Hoffman et al., 1999; Culnan & Armstrong, 1999). In turn, lower privacy concerns are likely to produce more favorable perceptions of e-tailer trustworthiness (Okazaki et al., 2009). At the same time, a privacy statement may serve as a signal of the firm's concern with its customers' well-being thus also having a positive impact on perceived e-tailer trustworthiness (Pan & Zinkhan, 2006). Hence, we posit the following hypotheses:

H1: Consumers' perceptions of the e-tailer privacy policy have a positive influence on their perceptions of the e-tailer trustworthiness.

H2: Consumers' perceptions of the e-tailer privacy policy have a negative influence on their privacy concerns.

H3: Consumers' privacy concerns have a negative influence on their perceptions of the etailer trustworthiness.

H4: The effect of consumers' perceptions of privacy policy on their perceptions of the e-tailer trustworthiness is partially mediated by consumers' privacy concerns.

Online Shopping Experience

General online shopping experience is likely to influence consumers' privacy concerns. To begin with, novice online shoppers have limited knowledge of the industry information practices, causing greater anxiety over their information privacy (Hoffman et al., 1999). In addition, research suggests that even experienced online shoppers tend to overestimate their knowledge of the Internet technology, including e-tailers' use of cookies to monitor customers' shopping behavior. For example, Jensen et al. (2005) found that 90.3% of experienced Internet users exhibited high confidence in their knowledge of cookies while only 15.5% of those making claims could actually demonstrate some simple cookie knowledge. With higher perceived knowledge of Internet technology, experienced shoppers may be less concerned with the e-tailer's ability to monitor their online behaviors because their knowledge of the Internet technology is already incorporated in their expectations about their online activity (Miyazaki, 2008). In contrast, lower perceived knowledge of inexperienced Internet users may result in heightened attention to different signs or signals of information security.

The above discussion also implies that online shopping experience helps shoppers to develop a general knowledge structure of typical online privacy protocols. Theoretically, increased familiarity leads to better knowledge structures or "schema" that include evaluative criteria and rules used in assessing new information (Marks & Olson, 1981). In the context of privacy statements, more experienced online shoppers are likely to rely on their existing schema in evaluating new privacy statements and determining their adequacy. Consequently, more experienced online shoppers are likely to have greater confidence in their evaluations of a new privacy statement than less experienced online shoppers who lack any definite evaluative criteria. Hence we hypothesize that online shopping experience will amplify the effect of consumer perceptions of the e-tailer privacy policy on their privacy concerns.

H5: Consumers' online shopping experience has a negative influence on their privacy concerns.

H6: An interaction effect between privacy policy perceptions and online shopping experience magnifies the relationship between privacy policy perceptions and privacy concerns.

METHOD AND RESULTS

Sample and Procedure

Survey respondents included undergraduate business students who received extra credit for their participation in this research. The student sample was deemed appropriate for this study because most online purchases are made by college-age consumers (Clemente, 1998) and other published studies have also used student subjects in testing theory-driven models of online behavior (Huang et al., 2004). A total of 280 students from six undergraduate marketing classes in southeastern United Stated completed the survey. However, after the missing data analysis the dataset was reduced to 271 respondents consisting of 133 males and 138 females. In terms of respondents' online shopping behaviors, they ranged from purchasing in multiple product categories (e.g., clothing, travel, electronics, etc.) and from multiple websites to limited product categories and one or two websites. The frequency of online shopping also varied, with most respondents (around 70 percent) making online purchases at least once a month.

The purchasing task in the survey involved online booking of air travel for an upcoming spring break. Online booking of air travel was chosen because of higher perceived privacy risk associated with this type of transaction (Bart et al., 2005). The survey participants received a questionnaire packet containing a print out of the homepage of a fictitious online booking agent, a copy of the agent's privacy policy, and the questionnaire. In the purchasing scenario, the fictitious booking agent was made to look like any other online travel agent (e.g., Expedia, Travelocity) to make sure that it was believable. It was described as a new website with great bargains on air travel. The booking process required customers to create an account where they had to respond to such personal questions as name and address, credit card information, and travel preferences (destinations, lodging, car rentals, and recreation). The scenario also referred survey participants to the agent's privacy policy, which was created to look very similar to the privacy policies of Expedia and Travelocity. Survey participants had ten minutes to review the materials in the questionnaire packet. Afterwards, the survey administrator collected the scenario materials leaving the respondents only with the questionnaire that they were required to complete in relation to their booking of air travel on the featured website. The questionnaire contained only the survey questions and did not require the participants to provide any personal information--i.e., the task of creating an account was hypothetical and not a requirement. The purpose of this procedure was to ensure that the participants would respond to survey questions from memory instead of referring back to the materials in the packet. Also, using a fictitious online booking agent helped us to avoid the confounding effects of website-specific experience and allowed us to focus on general shopping experience as the study intended.

Construct Measures

Table 1 presents construct measures used in the study. We developed our measures by translating theoretical definitions of the constructs into their operational definitions and then subjecting them to several rounds of pretests using a different sample of student respondents. Our goal was to develop valid and reliable measures that would allow us to estimate a series of structural models to test our hypotheses. Therefore, each construct was analyzed using exploratory factor analysis (EFA) prior to its inclusion in a measurement model, where it was further purified following the confirmatory factor analysis procedure (CFA). This section provides a brief summary of EFA results and reliability estimates (Cronbach's alpha) of the measures. CFA results are reported in the results section.

E-tailer trustworthiness was measured with four seven-point Likert-type scales designed to assess respondents' perceptions of the booking agent's dependability, competence, integrity, and responsiveness to customer needs. The EFA produced a single-factor solution explaining 71.03 percent of variance. Cronbach's alpha of the scale was 0.92.

Privacy concerns measures were adapted from Smith et al. (2006) scale of privacy concerns with some changes in the wording to make them more context-specific. The original scale is comprised of four subscales--collection, errors, unauthorized secondary use, and improper access--that measure general consumer concerns with privacy online. For the purpose of our study, we created items that closely resemble the items in the collection and unauthorized secondary use subscales in Smith et al. instrument. Thus, in our study privacy concerns were measured with six items addressing both secondary use of information and shopping anonymity dimensions of information privacy (Goodwin, 1991; Hoffman et al., 1999). Specifically, respondents were asked to indicate their level of agreement with six statements measuring their confidence in certain e-tailer behaviors that are directly related to protecting shoppers' information privacy. These responses were then reverse-coded to get the measures of privacy concerns. Thus a strong agreement with a statement: "When booking air travel on this website, I feel confident that this online booking agent would not sell my personal information to other companies without my knowledge" translated into a low privacy concern score for this statement. The EFA single-factor solution explained 80.37 percent of variance. Cronbach's alpha of the scale was 0.95.

Respondents' perceptions of the agent's information privacy policy were measured with four seven-point Likert-type scales that asked to recall whether the agent's privacy policy was available, clear, easy to understand, and could be considered credible. The EFA single-factor solution explained 70.63 percent of variance. Cronbach's alpha of the scale was 0.86.

To measure web-shopping experience, we asked respondents to indicate how long they had been shopping on the Internet, how often they made purchases on the Internet, and how they rated their knowledge of Internet shopping. These items were seven-point Likert-type scales. The EFA single-factor solution explained 76.48 percent of variance. Cronbach's alpha of the scale was 0.85.

Measurement Model

The proposed hypotheses were tested by estimating a series of models with covariance structure modeling, following a two-step approach (Anderson & Gerbing, 1988). Initially, a series of CFA models were estimated to ensure that all constructs had acceptable measurement properties. These models were consecutively estimated after being assessed in terms of fit, item loadings, and modification indices. In addition to chi-square, model fit was evaluated with the Comparative Fit Index (CFI), the Tucker-Lewis Index (TLI or NNFI), and the root mean square error of approximation (RMSEA). Values of .90 and above for CFI and TLI and values of .80 and less for RMSEA have been typically used as indicators of acceptable model fit (Browne & Cudeck, 1993; Hu & Bentler, 1995).

Prior to model estimation, however, all construct measures were centered by having their raw scores replaced with deviation scores (i.e., deviation score = variable score--variable mean). This procedure reduces the inherent multicollinearity between the interacting variables (Ping,

2003). Then we created an interaction term of privacy policy and online shopping experience by following Kenny and Judd (1984) procedure. According to this technique, the interaction term is specified using indicators that are the unique cross products of the two constructs (also see Ping, 2006 for a detailed discussion of latent variable interaction techniques). All measurement models discussed here were estimated using centered construct measures and included the newly created interaction term with its product indicators.

The initial measurement model with 29 manifest indicators and five latent constructs had unacceptable fit ([chi square] = 2054.16 (df = 367), CFI = .74, TLI = .71, and RMSEA = .12). After a careful examination of item loadings and modification indices, four indicators of the interaction term were dropped from the second estimation of the measurement model. The second model had a better yet still unacceptable fit and more items in the interaction term were dropped for their poor measurement properties. In sum, the measurement model was re-estimated three times, showing marked improvements in the model fit with each re-estimation. The final 21-item measurement model had a very good fit ([chi square] = 391.81 (df = 179), CFI = .95, TLI = .94, and RMSEA = .065). The interaction term retained four items, which is acceptable considering that some established techniques for testing interactions with structural equation modeling use only a subset of Kenny and Judd's (1984) product indicators (Jaccard & Wan, 1995; Marsh et al., 2004). Table 2 reports CFA factor loadings and error variances of the retained individual items while Table 3 provides construct correlations, average variance extracted, and internal consistency estimates.

As shown in Table 3, Cronbach's alpha, the measure of internal consistency, ranged from .85 to .95. Discriminant validity was assessed by comparing the square of the correlation (phi-square) between two constructs and their average variance extracted (AVE) estimates (Fornell & Larcker, 1981). Discriminant validity is supported when phi-square is less than the average AVE. This is the most stringent test of discriminant validity and was met for all possible pairs of the constructs. In sum, overall results indicated a good fit for the measurement model.

Mediation Analysis

Testing for mediation using structural equation modeling required estimating three structural models in order to establish the existence of a relationship between the exogenous and the endogenous variables (H1) and to meet the three mediation criteria (Baron & Kenny, 1986; Judd & Kenny, 1981): 1) the exogenous variable must affect the possible mediating variable (H2); 2) the mediator variable must affect the endogenous variable (H3), and 3) if the first two conditions are met and the mediating variable is controlled for, the effect between the exogenous and the endogenous variables must be dramatically reduced or non-existent (Brown, 1997). That is, a reduced effect between privacy policy perceptions and trustworthiness when controlling for privacy concerns would provide evidence in support for partial mediation as specified in H4. The fit of each model, path estimates and variance extracted of endogenous variables (including the mediator) are discussed in the following sections.

Model 1

The first structural model was estimated to determine the existence of relationships among privacy policy (exogenous variable), trustworthiness (endogenous variable) and privacy concerns (mediator). An additional path from online shopping experience to privacy concerns was estimated to test the hypothesis (H5) that posited a negative relationship between these two variables. All fit indices for this structural model indicated a good fit ([chi square] = 214.37 (df = 115), CFI = .97, TLI = .97, and RMSEA = .057) and all paths were significant and in predicted direction. Thus privacy policy appeared to have a significant negative effect on privacy concerns ([[gamma].sub.11] = -.46, t = -4.9) and a significant positive effect on trustworthiness ([[gamma].sub.21] = .66, t = 7.49). In addition, online shopping experience had a significant negative effect on privacy concerns (y12 = -.20, t = -3.24), providing support to H5. This model explained 19 percent of variance in privacy concerns and 28 percent of variance in trustworthiness. In sum, the results of this structural model met the first mediation criterion and produced evidence supporting H1 and H2.

Model 2

The main purpose of the second structural model was to establish a relationship between privacy concerns (mediator) and trustworthiness (endogenous variable), as prescribed by the second mediation criterion. The specification of this model was almost identical to Model 1, with the exception of two paths: the direct effect of privacy policy on trustworthiness was not estimated but an additional path from privacy concerns to trustworthiness was specified. The path from online shopping experience to privacy concerns was once again estimated to maintain the integrity of the mediation analysis. The fit indices of this structural model suggested a good fit ([chi square] = 246.79 (df = 115), CFI = .96, TLI = .96, and RMSEA = .063). The effect of privacy concerns on trustworthiness was significant and negative ([[beta].sub.21] = -.34, t = -5.97) thus meeting the second mediation criterion and providing support for H3. Similarly to Model 1, all other hypothesized effects were also significant and in predicted direction ([[gamma].sub.11] = -.44, t = -4.74; [[gamma].sub.12] = .21, t = -3.40). This model explained 18 percent of variance in privacy concerns and 14 percent of variance in trustworthiness. However, judging by the difference in Chi-square, Model 1 fit the data better than Model 2.

Model 3

The last model was estimated to test for partial mediation by privacy concerns as stated in H4. In this model, all three effects were specified: 1) from privacy policy to privacy concerns, 2) from privacy concerns to trustworthiness, and 3) from privacy policy to trustworthiness. If partial mediation exists, the effect of privacy policy on trustworthiness established in Model 1 should become smaller. As before, the effect of online shopping experience on privacy concerns was estimated to maintain consistency. The fit of this model was very good: [chi square] = 202.30 (df = 114), CFI = .98, TLI = .97, and RMSEA = .054. All path estimates were significant and in predicted direction ([[gamma].sub.21] = .54, t = 6.24; [[gamma].sub.11] = -.43, t = - 4.64; [[gamma].sub.12] = -.20, t = -3.33; [[beta].sub.21] = -.19, t = -3.48). Also, as expected, controlling for privacy concerns reduced the effect of privacy policy on trustworthiness, although this reduction was not substantial (from .66 to .54). However, the Chi-square difference test suggests that the partial mediation model is a little more parsimonious than the direct effects model (i.e., Model 1): [[chi square].sub.diff] = 12.07, [df.sub.diff] = 1. In sum, these results suggest that privacy policy has both direct and indirect effects on perceived e-tailer trustworthiness mediated by the customer's privacy concerns. In addition, the effect of online shopping experience on privacy concerns ([[beta].sub.21]) was consistently significant and negative in all three models, which offers support to H5. The partial mediation model explained 18 percent of variance in privacy concerns and 30 percent in trustworthiness.

Interaction Analysis

The last hypothesis (H6) predicted a positive interaction between privacy policy and online shopping experience. To test for the interaction effect, we estimated a model where privacy policy x online shopping experience interaction term was specified as a predictor of privacy concerns. All other effects were the same as in Model 3 (partial mediation model). The fit of the interaction model was good ([chi square] = 396.62 (df = 181), CFI = .95, TLI = .94, and RMSEA = .065) and, judging by the unstandardized path estimate, the effect of the privacy policy x online shopping experience interaction term was significant but small ([[gamma].sub.13] = .07, t = 2.14). Also, consistent with our hypothesis, the effect of the interaction term was positive suggesting that for more experienced online shoppers the e-tailer's information privacy policy is likely to have a stronger impact on their privacy concerns than for less experienced shoppers. This interaction model explained 19 percent of variance in privacy concerns and 30 percent of variance in trustworthiness. We will discuss these findings, their implications, and future research opportunities in the following section.

GENERAL DISCUSSION

Discussion of Results

This study complements and extends existing literature on information privacy and online trust in several ways. First, we draw a clear distinction between trust and trustworthiness, noting their conceptual distinctions that are likely to have important implications for the interpretation of the empirical results involving these two constructs. Second, we show that consumers' privacy concerns partially mediate the effect of information privacy policy on e-tailer trustworthiness. Our findings suggest that information privacy may play a dual role in shaping customers' perceptions of e-tailer trustworthiness: 1) indirectly--by informing customers about the intended uses for their personal information and thus reducing their privacy concerns and 2) directly--by serving as a signal of the e-tailer's integrity and general concern for customers' well-being. In addition, we emphasize the importance of considering consumers' experience with online shopping when studying their privacy perceptions online. In our study, more experienced online shoppers demonstrated lower privacy concerns and their perceptions of the agent's privacy policy had a stronger impact on their privacy concerns than the perceptions of less experienced online shoppers. These findings corroborate Huang et al.'s (2004) study where experienced online shoppers demonstrated lower risk perceptions associated with online shopping than novice or even non-shoppers (i.e., browsers). Furthermore, they suggest that more experienced online shoppers are likely to react more strongly to e-tailers' information practices than less experienced online shoppers because of their better developed "schema" of acceptable information practices. The moderating effect of online shopping experience also implies that less experienced online shoppers require additional assurances about the safety of providing personal information to a particular website. With higher privacy concerns and limited online shopping experience, these novices are likely to place greater trust in third party seals than in the website's privacy policy. This supposition offers an interesting opportunity for follow up research.

In sum, our study highlights the important role of privacy statements in reducing shoppers' privacy concerns and helping online companies to communicate their trustworthiness. Despite the fact that some e-tailers do not post their privacy policies online (Tang et al., 2008) and some privacy policies are not easy to comprehend (Vail et al. 2008), our findings provide clear evidence that information privacy policies can effectively mitigate online shoppers' privacy concerns and enhance their perceptions of e-tailer trustworthiness. Information privacy policy is not only a signal of the website's integrity but also a form of social contract that promises shoppers that their privacy will be protected. Therefore, it must occupy a prominent space on any website to make sure that customers are always aware of its existence.

Future Research Avenues

There are still many things we don't know about consumer attitudes toward their privacy online. For example, which factors influence the relative effectiveness of information privacy policies? Pan and Zinkhan (2006) found that online shoppers prefer short and more comprehensible privacy statements, but information privacy policy presentation format (in terms of wording) does not affect consumers' perceptions of e-tailer trustworthiness. Also, how much do consumers truly value their information privacy? What is the trade-off between consumers' desire for information privacy and their economic self-interest, including their desire for personalized market offerings? Some research suggests that consumers could be incentivized to provide their personal information (Ward et al., 2005). Another research avenue would be to investigate how information privacy breach of one website affects online shoppers' privacy concerns and their propensity to pay closer attention to privacy policies of individual websites. As Van Slyke et al. (2006) suggest, consumers' privacy concerns are general in nature and apply to the entire online marketspace as a whole. Hence, a single violation of consumer privacy (e.g., new Facebook features that jeopardize user privacy) could potentially plant a seed of skepticism and increase consumers' privacy concerns, making online shoppers less trusting of information privacy policies.

Limitations

Like any research, our study is not without limitations. One of these limitations involves our data collection method. The survey method cannot replicate the actual experience of booking air travel online. The interactivity of online shopping and the actual necessity to create an account and provide personal information will certainly elicit consumer thought processes that cannot be evoked with just a visual stimulus of an online booking agent and a scenario, regardless of how vivid and descriptive they may be. This may explain why our findings were relatively weak, although statistically significant.

Another limitation concerns our use of college students as survey respondents. Although students were appropriate for this study given their familiarity with online shopping and the nature of the task, our findings cannot be generalized beyond student population. Generally speaking, college students are likely to be better educated and more comfortable with information technology than their non-college educated counterparts.

Despite these limitations, however, our study makes a noteworthy contribution to published research on information privacy and online trust. Here we showed a process by which information privacy policy affects online shoppers' perceptions of e-tailer trustworthiness and demonstrated the importance of considering online shoppers' experience in assessing the effectiveness of a website's information privacy policy.

REFERENCES

Anderson, J. C. & Gerbing, D. W. (1988). Structural equation modeling in practice: a review and recommended two-step approach. Psychological Bulletin, 103 (3), 411-423.

Baron, R. M. & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51 (6), 1173-1182.

Bart, Y., Shankar, V., Sultan, F., & Urban, G. L. (2005). Are the drivers and role of online trust the same for all web sites and consumers? A large-scale exploratory empirical study. Journal of Marketing, 69 (4), 133-152.

Bhattacherjee, A. (2002). Individual trust in online firms: scale development and initial test. Journal of Management Information Systems, 19 (1): 211-241.

Brown, R. L. (1997). Assessing specific mediational effects in complex theoretical models. Structural Equation Modeling, 4 (2), 142-156.

Browne, M. W. & Cudeck, R. (1993). Alternative ways of assessing model fit. In A. Bollen & J. S. Long (Eds.), Testing Structural Equation Models (136-162). Newbury Park, CA: Sage Publications.

Campbell, A. J. (1997). Relationship marketing in consumer markets. Journal of Direct Marketing, 11 (3), 44-57.

Clemente, P. (1998). The state of the net: The new frontier. New York, NY: McGraw-Hill.

Cranor, L. F., Guduru, P. & Arjula, M. (2006). User interfaces for privacy agents. ACM Transactions on Computer-Human Interaction, 13 (2), 135-178.

Culnan, M. J. (1995). Consumer awareness of name removal procedures: implications for direct marketing. Journal of Direct Marketing, 9 (2), 10-19.

Culnan, M. J. & Armstrong, P. K. (1999). Information privacy concerns, procedural fairness, and impersonal trust: an empirical investigation. Organization Science, 10 (1), 104-115.

Doney, P. M. & Canon, J. P. (1997). An examination of the nature of trust in buyer-seller relationships. Journal of Marketing, 61 (2), 35-51.

Dunfee, T. W., Smith, N. C. & Ross, W. T. (1999). Social contracts and marketing ethics. Journal of Marketing, 63 (3), 14-32.

Eastlick, M. A., Lotz, S. L. & Warrington, P. (2006). Understanding online b-to-c relationships: an integrated model of privacy concerns, trust, and commitment. Journal of Business Research, 59 (8), 877-886.

Fornell, C. & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18 (1), 39-50.

Gambetta, D. (1988). Trust: Making and breaking cooperative relationships. New York, NY: Blackwell.

Gefen, D., Karahanna, E. & Straub, D. W. (2003). Trust and TAM in online shopping: an integrated model. MIS Quarterly, 27 (1), 51-90.

Gengler, C. E. & Leszczyc, P. (1997). Using customer satisfaction research for relationship marketing: a direct marketing approach. Journal of Direct Marketing, 11 (4), 36-41.

Goodwin, C. (1991). Privacy: recognition of a consumer right. Journal of Public Policy and Marketing, 10 (1), 149166.

Hoffman, D. L., Novak, T.P. & Peralta, M. A. (1999). Information privacy in the marketspace: implications for the commercial uses of anonymity on the web. The Information Society, 15 (2), 129-140.

Hu, L. & Bentler, P. M. (1995). Evaluating model fit. In R. H Hoyle (Ed.), Structural Equation Modeling: Concepts, Issues, and Applications (76-99). Thousand Oaks, CA: Sage Publications.

Huang, W., Schrank, H. & Dubinsky, A. J. (2004). Effect of brand name on consumers' risk perceptions of online shopping. Journal of Consumer Behavior, 4 (1), 40-50.

Jaccard, J. & Wan, C. K. (1995). Measurement error in the analysis of interaction effects between continuous predictors using multiple regression-multiple indicator and structural equation approaches. Psychological Bulletin, 117 (2): 348-357.

Jensen, C., Potts, C. & Jensen, C. (2005). Privacy practices of internet users: self-reports versus observed behavior. International Journal of Human-Computer Studies, 63 (1-2), 203-227.

Judd, C. M. & Kenny, D. A. (1981). Process analysis: estimating mediation in treatment evaluations. Evaluation Review, 5 (5): 602-619.

Kenny, D. A. & Judd, C. M. (1984). Estimating the nonlinear and interactive effects of latent variables. Psychological Bulletin, 96 (1), 201-210.

Malhotra, N. K., Kim, S. S. & Agarwal, J. (2004). Internet users information privacy concerns (IUIPC): the construct, the scale, and a causal model. Information Systems Research, 15 (4), 336-355.

Marks, L. & Olson, J.(1981). Towards a cognitive structure conceptualization of product familiarity. In K. B.

Monroe (Ed.), Advances in Consumer Research (145-150). Ann Arbor, MI: Assoc.for Consumer Research.

Marsh, H. W., Wen, Z. & Hau, K. (2004). Structural equation models of latent interactions: evaluation of alternative estimation strategies and indicator construction. Psychological Methods, 9 (3), 275-300.

McKnight, H. D., Choudhury, V. & Kacmar, C. (2002). The impact of initial consumer trust on intentions to transact with a web site: a trust building model. The Journal of Strategic Information Systems, 11 (3-4), 297-323.

Meinert, D. B., Peterson, D. K., Criswell, J. R. & Crossland, M. D. (2006). Privacy policy statements and consumer willingness to provide personal information. Journal of Electronic Commerce in Organizations, 4 (1), 1-17.

Milne, G. R., & Gordon, M. E. (1993). Direct mail privacy-efficiency trade-offs within an implied social contract framework. Journal of Public Policy and Marketing, 12 (2), 206-215.

Milne, G. R. & Culnan, M. J. (2004). Strategies for reducing online privacy risks: why consumers read (or don't read) online privacy notices. Journal of Interactive Marketing, 18 (3), 15-29.

Milne, G. R., Culnan, M. J. & Greene, H. (2006). A longitudinal assessment of online privacy notice readability. Journal of Public Policy & Marketing, 25 (2), 238-249.

Miyazaki, A. D. (2008). Online privacy and the disclosure of cookie use: effects on consumer trust and anticipated patronage. Journal of Public Policy and Marketing, 27 (1), 19-33.

Nehf, J. P. (2007). Shopping for privacy on the internet. Journal of Consumer Affairs, 41 (2), 351-375.

Okazaki, S., Li, H. & Hirose, M. (2009). Consumer privacy concerns and preference for degree of regulatory control. Journal of Advertising, 38 (4), 63-77.

Pan, Y. & Zinkhan, G. M. (2006). Exploring the impact of online privacy disclosures on consumer trust. Journal of Retailing, 82 (4), 331-338.

Ping, R. A. (2003). Latent Variable Interactions and Quadratics in Survey Data: A Source Book for Theoretical Model Testing (Second Edition). [On-line monograph]. Available at http://www.wright.edu/~robert.ping/intquad2/toc2.htm [accessed on 12 May 2010].

Ping, R. A. (2006). Frequently asked questions (FAQ's) about interactions and quadratics. [On-line paper]. Available at http://www.wright.edu/~robert.ping/research1.htm [accessed on 12 May 2010].

Proctor, R. W., Ali, M. A. & Vu, K. L. (2008). Examining usability of web privacy policies. International Journal of Human-Computer Interaction, 24 (3), 307-328.

Rifon, N. J., LaRose, R. & Choi, S. M. (2005). Your privacy is sealed: effects of web privacy seals on trust and personal disclosures. Journal of Consumer Affairs, 39 (2), 339-362.

Rousseau, D. M., Bitkin, S. B., Burt, R. S. & Camerer, C. (1998). Not so different after all: a cross-discipline view of trust. Academy of Management Review, 23 (3), 393-404.

Smith, H. J., Milberg, S. J. & Burke, S. J. (1996). Information privacy: measuring individuals' concerns about organizational practices. MIS Quarterly, 20 (2), 167-196.

Tang, Z., Hu, Y. & Smith, M. D. (2008). Gaining trust through online privacy protection: self-regulation, mandatory standards or caveat emptor. Journal of Management Information Systems, 24 (4), 153-173.

Vail, M. W., Earp, J. B. & Anton, A. L. (2008). An empirical study of consumer perceptions and comprehension of web site privacy policies. IEEE Transactions on Engineering Management, 55 (3), 442-454.

Van Slyke, C. J., Shim, T., Johnson, R. & Jiang, J. (2006). Concern for information privacy and online consumer purchasing. Journal of the Association for Information Systems, 7 (6), 415-443.

Ward, S., Bridges, K. & Chitty, B. (2005). Do incentives matter? An examination of on-line privacy concerns and willingness to provide personal and financial information. Journal of Marketing Communications, 11 (1), 21-40.

Zeithaml, V. A., Parasuraman, A. & Malhotra, A. (2002). Service quality delivery through web sites: a critical review of extant knowledge. Journal of the Academy of Marketing Science, 30 (4), 362-375.

Elena Kiryanova Bernard, University of Portland

Igor Makienko, University of Nevada Reno

Table 1. Construct Measures

E-tailer Trustworthiness

When it comes to booking air travel on this website, I feel
that this agent is:

1). Very Undependable (1)/Very Dependable (7)

2). Very Incompetent (1)/Very Competent (7)

3). Of Very Low Integrity (1) / Of Very High Integrity (7)

4). Very Unresponsive to Customer Needs (1)/Very Responsive
to Customer Needs

Information Privacy Concerns

When it comes to booking air travel on this website, I feel confident
That ...

1). This online booking agent would not "spy" on me when I surf the
Internet (Strongly Disagree/Strongly Agree).

2). This online booking agent would not sell my personal information
to other companies without my knowledge (Strongly

Disagree/Strongly Agree).

3). This online booking agent would not disclose my personal
information to other parties without my permission (Strongly
Disagree/Strongly Agree).

4). This online booking agent would not track my shopping habits
or purchases on other websites without my knowledge
(Strongly Disagree/Strongly Agree).

5). his online booking agent would request my permission before
disclosing my personal information to other parties
(Strongly

Disagree/Strongly Agree).

6). This online booking agent would ask my permission before tracking
my surfing behaviour on the Internet (Strongly
Disagree/Strongly Agree).

Privacy Policy

This online booking agent's information privacy policy is...
(Not Available/Available), (Difficult to Understand/Easy
to Understand), (Confusing/Clear), (Not at all Credible/Very Credible).

Online Shopping Experience

1). How long have you been shopping on the Internet? (Just Started/Have
Been Shopping for a Very Long Time)

2). How often do you make purchases on the Internet? (Very Rarely
/Very Frequently)

3). How would you rate your knowledge of Internet shopping? (Know
Very Little/Know Everything About It)

Table 2: CFA Factor Loadings and Error Variances
(Final Measurement Model)

 Items           Completely Standardized Loadings (error variances)

              PP          OSE        INTER        IPC         ET

PP1        .68 (.53)
PP2        .87 (.24)
PP3        .77 (.40)
PP4        .78 (.39)
OSE1                   .78 (.39)
OSE2                   .87 (.25)
OSE3                   .77 (.41)
INTER1                             .93 (.13)
INTER2                             .67 (.55)
INTER3                             .76 (.42)
INTER4                             .83 (.31)
IPC1                                           .81 (.34)
IPC2                                           .90 (.20)
IPC3                                           .88 (.22)
IPC4                                           .90 (.19)
IPC5                                           .88 (.23)
IPC6                                           .88 (.22)
ET1                                                        .91 (.18)
ET2                                                        .91 (.17)
ET3                                                        .90 (.19)
ET4                                                        .90 (.19)

Note: PP--Privacy Policy; OSE--Online Shopping Experience; INTER--
PP x OSE interaction term; IPC--Information Privacy Concerns, ET--
E-tailer Trustworthiness.

Table 3: Correlations, Reliabilities and Average Variance Extracted
Estimates

            Constructs                Alpha     AVE       Correlations

                                                          PP      OSE

Privacy Policy (PP)                    .86      .61      1.00
Online Shopping Experience (OSE)       .85      .65      .18      1.00
Interaction Term (PP x OSE)            .88      .65     .07 *    .02 *
Information Privacy Concerns (IPC)     .95      .77      -.36     -.28
E-tailer Trustworthiness (ET)          .95      .82      .51      .26

            Constructs                      Correlations

                                      PPxOSE    IPC       ET

Privacy Policy (PP)
Online Shopping Experience (OSE)
Interaction Term (PP x OSE)            1.00
Information Privacy Concerns (IPC)    .10 *     1.00
E-tailer Trustworthiness (ET)         .02 *     -.37     1.00

Note: * correlation is not significant at p = .05 level.